Skip to main content
Tessa NewsLink
Paste a health news link, or browse
Latest checks

Source study found

Story checked

Q&A: AI tool automates echocardiogram assessment of the heart's right ventricle (opens in a new tab)

medicalxpress.com · 2026-10-20

Short answerEvidenceSource

Short answer

Mostly not supported

Mostly not supported.

The claims we could check match the study, but some claims were not covered by the evidence reviewed.

  • 2 supported
  • 5 not covered

Checked against the study summary. The full text wasn't available, so some details couldn't be settled either way.

Share this check

Follow the evidence trail
1
2

NewsLink checks it

Mostly not supported

Every claim we could check holds up. Two of seven claims match the study. This overall rating is based only on the claims we could check. Five claims the study doesn't address.

  • 2 supported
  • 5 not covered
Open claim evidence
3
Then inspect each claim

Evidence layer

Claim by claim

Each claim gets a verdict. Expand it to see the evidence directly below.

7 claims in this story

Showing all 7 claimsChoose a verdict to focus the list.

Claim 1 of 7Not coveredUsing more than 380,000 echocardiogram video clips from about 17,000 patients, the researchers trained the model and tested it in about 9,100 primary care patients plus an independent dataset of 831 patients from Hungary.View evidenceHide evidence

As statedmore than 380,000 echocardiogram video clips; about 17,000 patients; about 9,100 primary care patients; 831 patients

Why this verdict

The abstract supports the main dataset counts: 17,141 ambulatory cardiology patients, 51,188 studies, 383,881 video clips; validation in 9,112 primary care patients; and a 831-patient cohort with 3D TTE-derived RV ejection fraction. However, the supplied abstract-level profile does not verify that the 831-patient dataset was from Hungary, so that part cannot be confirmed at this evidence depth.

1

Study evidence

DROID-RV was trained on a large ambulatory cardiology echocardiography dataset (17,141 patients; 51,188 studies; 383,881 video clips).

“We trained DROID-RV in 17,141 ambulatory cardiology patients (51,188 studies and 383,881 video clips)”
2

Study evidence

In an external primary care cohort of 9,112 patients, DROID-RV estimated RV end-diastolic diameter with reported accuracy across test samples.

“validated model predictions in 9,112 primary care patients”
Claim 2 of 7Not coveredDROID-RV estimated right ventricular diameter to within about 3 millimeters on average and identified reduced right ventricular function with high accuracy, including in the outside data set.View evidenceHide evidence

As statedabout 3 millimeters on average; high accuracy

Why this verdict

The abstract supports qualitative accuracy across multiple test samples and external/reference validation, while also noting increased prediction error at extremes of RV size and function. It does not provide the claimed numeric average error of about 3 mm, nor detailed accuracy metrics for dysfunction classification, so the quantitative version of the claim is not verifiable from the abstract profile.

1

Study evidence

DROID-RV was trained on a large ambulatory cardiology echocardiography dataset (17,141 patients; 51,188 studies; 383,881 video clips).

“We trained DROID-RV in 17,141 ambulatory cardiology patients (51,188 studies and 383,881 video clips)”
2

Study evidence

In an external primary care cohort of 9,112 patients, DROID-RV estimated RV end-diastolic diameter with reported accuracy across test samples.

“validated model predictions in 9,112 primary care patients”
Claim 3 of 7Not coveredModel predictions of larger right ventricular size or worse right ventricular function were associated with heart failure, atrial fibrillation, and mortality; patients with predicted dilated right ventricles were twice as likely to die within five years compared with patients with predicted normal right ventricular size.View evidenceHide evidence

As statedtwice as likely to die within five years

Why this verdict

The abstract supports that DROID-RV-derived RV size and function predictions were associated with incident mortality, atrial fibrillation, and heart failure. It does not provide the claimed effect magnitude, five-year time horizon, or specific comparison that patients with predicted dilated RVs were twice as likely to die, so those details are not verifiable at abstract depth.

1

Study evidence

DROID-RV–derived predictions of RV size and RV systolic dysfunction were associated with higher incidence of mortality, atrial fibrillation, and heart failure on follow-up.

“We additionally show that RV size and function predictions are associated with incident mortality, atrial fibrillation and heart failure.”
Claim 4 of 7Not coveredThe article says the model could help echocardiographers read echocardiograms faster, flag abnormal exams, and identify patients who may benefit from additional MRI testing.View evidenceHide evidence

Why this verdict

The abstract supports only broad proposed uses: automating RV assessment and enabling risk stratification at scale. It does not verify workflow claims about reading echocardiograms faster, flagging abnormal exams, or selecting patients for additional MRI testing. These may be discussed outside the abstract, but they are not verifiable from the supplied abstract-level profile.

1

Study evidence

DROID-RV was trained on a large ambulatory cardiology echocardiography dataset (17,141 patients; 51,188 studies; 383,881 video clips).

“We trained DROID-RV in 17,141 ambulatory cardiology patients (51,188 studies and 383,881 video clips)”
2

Study evidence

DROID-RV–derived predictions of RV size and RV systolic dysfunction were associated with higher incidence of mortality, atrial fibrillation, and heart failure on follow-up.

“We additionally show that RV size and function predictions are associated with incident mortality, atrial fibrillation and heart failure.”
Claim 5 of 7Not coveredThe researchers say predicted right ventricular dysfunction remained independently associated with mortality and atrial fibrillation even after accounting for the echocardiographer's own assessment, suggesting the model may capture prognostic information beyond the measurements it was trained to reproduce.View evidenceHide evidence

Why this verdict

The abstract supports associations between model-derived RV predictions and incident mortality, atrial fibrillation, and heart failure, but it does not report adjustment for the echocardiographer’s own assessment or demonstrate incremental prognostic value beyond the measurements the model was trained to reproduce. The claim may be true in the full paper, but it is not verifiable at abstract depth.

1

Study evidence

DROID-RV–derived predictions of RV size and RV systolic dysfunction were associated with higher incidence of mortality, atrial fibrillation, and heart failure on follow-up.

“We additionally show that RV size and function predictions are associated with incident mortality, atrial fibrillation and heart failure.”
Claim 6 of 7SupportedA new study develops and tests DROID-RV, a deep learning model that estimates right ventricular size and function from echocardiograms and assesses how those estimates relate to later health outcomes.View evidenceHide evidence

Why this verdict

The abstract supports that the study developed/trained DROID-RV to estimate right ventricular end-diastolic diameter and identify RV systolic dysfunction from 2D transthoracic echocardiography videos, validated it in external/reference cohorts, and assessed associations of model-derived RV measures with incident outcomes. This is supported as a development/validation and prognostic-association claim, not as evidence that RV measures cause those outcomes.

1

Study evidence

DROID-RV was trained on a large ambulatory cardiology echocardiography dataset (17,141 patients; 51,188 studies; 383,881 video clips).

“We trained DROID-RV in 17,141 ambulatory cardiology patients (51,188 studies and 383,881 video clips)”
2

Study evidence

In an external primary care cohort of 9,112 patients, DROID-RV estimated RV end-diastolic diameter with reported accuracy across test samples.

“validated model predictions in 9,112 primary care patients”
Claim 7 of 7SupportedThe study asked whether a deep learning model could automatically estimate right ventricular size and detect reduced pumping function directly from routine echocardiogram videos, and whether the automated estimates could help predict death, atrial fibrillation, or heart failure.View evidenceHide evidence

Why this verdict

The abstract states that DROID-RV can estimate RV end-diastolic diameter and identify RV systolic dysfunction from transthoracic echocardiography videos, and that model-derived RV size/function predictions were associated with incident mortality, atrial fibrillation, and heart failure. The claim’s 'help predict' framing is broadly consistent with the paper’s prognostic association/risk-stratification framing at abstract depth.

1

Study evidence

DROID-RV was trained on a large ambulatory cardiology echocardiography dataset (17,141 patients; 51,188 studies; 383,881 video clips).

“We trained DROID-RV in 17,141 ambulatory cardiology patients (51,188 studies and 383,881 video clips)”
2

Study evidence

In an external primary care cohort of 9,112 patients, DROID-RV estimated RV end-diastolic diameter with reported accuracy across test samples.

“validated model predictions in 9,112 primary care patients”
Then look for missing context

Context layer

What the story left out

Important study details the story did not include.

  • Prediction error increased at extremes of RV size and function.

    This is a material limitation in the abstract because it affects interpretation of model accuracy, especially for abnormal or extreme measurements. The story caveats supplied do not mention this limitation.

    From secondary_data; deep learning model training; external validation / secondary-data testing; comparative validation again

  • The abstract does not provide numeric performance metrics for model accuracy, classification performance, or 3D-reference validation.

    The story reports specific performance language including an average error of about 3 mm and 'high accuracy,' but the abstract-level paper profile only supports qualitative accuracy and explicitly notes lack of numeric performance metrics.

    From secondary_data; deep learning model training; external validation / secondary-data testing; comparative validation again

  • The abstract does not report effect sizes, confidence intervals, follow-up duration, statistical model details, adjustment covariates, or outcome ascertainment procedures for the prognostic analyses.

    This limitation is important because the story includes a specific mortality magnitude and five-year time frame, and mentions adjustment variables, while the abstract-level profile says these details are not provided in the abstract.

    From Observational prognostic association analysis

3 things the story did carry across
  • DROID-RV was developed/trained to estimate RV end-diastolic diameter and identify RV systolic dysfunction from 2D transthoracic echocardiography videos, using 17,141 ambulatory cardiology patients, 51,188 studies, and 383,881 video clips.
  • The model was validated in 9,112 primary care patients and in 831 patients with RV ejection fraction assessed by 3D transthoracic echocardiography.
  • DROID-RV-derived RV size and function predictions were associated with incident mortality, atrial fibrillation, and heart failure in observational prognostic analyses.
Then read the study layer

Study layer

Study at a glance

Scan the study first. Expand only the parts you want to inspect.

Pieces of work

4

Evidence read

study summary

Lead result

secondary data

1Lead resultsecondary dataDevelop and train a deep learning model (DROID-RV) to estimate RV end-diastolic diameter and detect RV systolic dysfunction from 2D transthoracic echocardiography videos at scale.secondary data; deep learning model trainingExpand

In plain English

The paper presents DROID-RV, a deep learning model trained on 17,141 ambulatory cardiology patients (51,188 studies; 383,881 video clips) from 2‑D transthoracic echocardiography to (1) estimate right ventricular end-diastolic diameter (regression) and (2) identify RV systolic dysfunction (classification). The abstract reports that the model accurately estimates RV end-diastolic diameter and identifies RV dysfunction across multiple test samples, but that prediction error increases at extremes of RV size and function. The abstract does not provide details on model architecture, training hyperparameters, or the provenance of supervisory labels.

Key findings

  • DROID-RV was trained on a large ambulatory cardiology echocardiography dataset (17,141 patients; 51,188 studies; 383,881 video clips).
  • DROID-RV can estimate right ventricular end-diastolic diameter from 2-D transthoracic echocardiography videos.
“We trained DROID-RV in 17,141 ambulatory cardiology patients (51,188 studies and 383,881 video clips)”
What this piece can’t prove
  • Abstract does not specify how ground-truth labels for RV diameter or dysfunction were obtained (e.g., reader annotation, measurements, or reference modalities).

3 further details could not be confirmed from the summary.

2secondary dataValidate DROID-RV performance across external cohorts and against a higher-fidelity reference of RV function (3D transthoracic echocardiography-derived RV ejection fraction).external validation / secondary-data testingExpand

In plain English

The study externally validated the deep-learning model DROID-RV in a separate primary care cohort (9,112 patients) and against a higher-fidelity reference of RV function (831 patients with 3D transthoracic echocardiography-derived RVEF). DROID-RV, trained on 17,141 ambulatory cardiology patients (51,188 studies, 383,881 video clips), produced estimates of RV end-diastolic diameter and identified RV systolic dysfunction in external test samples; however, model prediction error increased at extremes of RV size and function. The abstract reports accuracy across multiple test samples but does not provide numeric performance metrics or detailed cohort characteristics.

Key findings

  • In an external primary care cohort of 9,112 patients, DROID-RV estimated RV end-diastolic diameter with reported accuracy across test samples.
  • DROID-RV identified RV systolic dysfunction and was validated against 3D transthoracic echocardiography-derived RV ejection fraction in 831 patients.
“validated model predictions in 9,112 primary care patients”
What this piece can’t prove
  • No details provided about primary care cohort demographics, case-mix, image acquisition differences, or how the 3D TTE subset was selected.
  • Unclear whether the external cohorts reflect populations and imaging protocols sufficiently different from the training set to fully assess generalizability.
  • Abstract notes increased prediction error at extremes but does not quantify the magnitude or clinical impact of this degradation.

1 further detail could not be confirmed from the summary.

3secondary dataValidate DROID-RV performance across external cohorts and against a higher-fidelity reference of RV function (3D transthoracic echocardiography-derived RV ejection fraction).comparative validation against 3D transthoracic echocardiography RVEFExpand

In plain English

Validation of DROID-RV predictions against RV ejection fraction measured by three-dimensional transthoracic echocardiography (3D TTE) in a distinct cohort of 831 patients. The model's 2D echocardiography–based predictions were compared/anchored to 3D TTE-derived RVEF; the abstract reports accurate estimation of RV end-diastolic diameter and identification of RV systolic dysfunction overall, with model prediction error increasing at extremes of RV size and function. The abstract does not report numeric performance metrics for this 3D-reference cohort or details of the pairing/analysis methods.

Key findings

  • When anchored to RV ejection fraction measured by 3D transthoracic echocardiography in 831 patients, DROID-RV (from 2D TTE videos) is reported to accurately estimate RV end-diastolic diameter and to identify RV systolic dysfunction.
“validated model predictions in ... 831 patients with RV ejection fraction assessed by three-dimensional transthoracic echocardiography”
What this piece can’t prove
  • The abstract notes increased error at extremes of RV size/function but does not quantify the magnitude or implications for clinical classification.

2 further details could not be confirmed from the summary.

4secondary dataDemonstrate clinical/prognostic relevance by testing whether DROID-RV-derived RV size/function predictions are associated with incident cardiovascular outcomes (mortality, atrial fibrillation, heart failure).Observational prognostic association analysisExpand

In plain English

The study reports that DROID-RV model–derived estimates of right ventricular (RV) size and RV systolic dysfunction are associated with incident clinical outcomes (mortality, atrial fibrillation, and heart failure) on follow-up, indicating prognostic relevance of the automated two-dimensional echocardiography measurements.

Key findings

  • DROID-RV–derived predictions of RV size and RV systolic dysfunction were associated with higher incidence of mortality, atrial fibrillation, and heart failure on follow-up.
“We additionally show that RV size and function predictions are associated with incident mortality, atrial fibrillation and heart failure.”
What this piece can’t prove
  • Increased model error at extremes of RV size/function may limit applicability of associations across the full range of values.

2 further details could not be confirmed from the summary.

Finally, the search trail

Method layer

NewsLink found the paper. Tessa takes you deeper.

NewsLink checks the story. Tessa is where you inspect the paper, authors, evidence, and research context.

Papers considered

The selected paper, plus nearby candidates.

PubMed, Europe PMC, Crossref · 36 candidate papers

Candidate

Al-Alusi dan sains modern: penafsiran Al-Alusi terhadap ayat kauniyah (Al-Zumar: 5) dalam kitab Tafsir Ruh al-Ma’ani

JPGI (Jurnal Penelitian Guru Indonesia) · 2026 · Crossref

Source (opens in a new tab)
Candidate

Tafsir Ruh al-Maʿani karya Al-Alusi: Analisis Biografis, Kontekstual, dan Metodologis

AL-IBANAH · 2026 · Crossref

Source (opens in a new tab)

And 30 more candidates considered.